Research On Key Technologies In The H.264/AVC

The newest video coding standard H.264/AVC is developed collaboratively by the ITU-T and ISO/IEC. The standard is designed to give a new technical solution for a broad range of applications, such as very low bit rate conversational services and entertainment quality broadcast, interactive video-on-demand services etc. By adopting a number of new coding techniques, H.264/AVC has achieved a significant improvement in compression performance and a"network-friendly"video representation relative to existing standards.In this dissertation, the new techniques in H.264/AVC are firstly discussed, experimental results prove the advantage of the new standard then.Secondly, we discuss the most important module in H.264/AVC——motion estimation. Based on the new characters of motion estimation about H.264/AVC, we develop a new algorithm called a fast motion estimation algorithm based on characters of motion, in which a novel method used to determine the initial motion vector search position is proposed based on the characters of H.264/AVC. The proposed algorithm is improved by incorporating a new motion determination criteria and an adaptive search strategy. The results show that the new algorithm achieves speedup with negligible PSNR drop . Furthermore, variable coding modes are employed in H.264/AVC,which makes the mode selection a time-consuming component. In this dissertation, a new algorithm called a fast mode decision algorithm based on mode reference aggregate is proposed. We first develop the concept of mode reference aggregate. Then based on the mode reference aggregate, a fast mode decision algorithm is proposed. This method can significantly reduce number of candidate mode,thus simplifying the procedure of mode decision.Finally, the error control technologies that are relatively often adopted in video communication systems are explained in detail. The error control tools with the overall system analysis in the H.264/AVC standard are discussed then. The work in this respect is carried to offer valuable materials for error control algorithms, and lay a good foundation further for future work.